The systems model and political science BY JAN-ERIK LANE

§0. In a provocative review and partly brilliant ana­lysis of the Easton critique H. Bang raises the basic question of the applicability of a systems model to the understanding of political data (Bang 1981). Bang shows convincingly that there are different approaches to the ontological as well as the epistemological implications of the relevance of the systems modeling of political data outlined by David Easton in his classical text A Systems Analysis of Political Life (1965). Bang's framework - complicated and not always quite exact - may be reduced to a simple 2 x 2 Table (p. 5).

To both the Understanding Tradition and the Critical Tradition the existential recognition that the social reality includes subjective phenomena like pieces of cognition, volition and valuation has profound consequences, but the Critical Tradition distinguishes itself by its insistance upon the fun­damentally subjective mode of social science or­ganisation, the emphasis on the interest orien­tation of knowledge and the denial of the pos­sibility of a value-neutral Erkenntnis. Bang fails to make the crucial distinction between two types of Explanatory Traditions: while both firmly be­

lieve in the possibility of making distinctions be­tween descriptive and prescriptive language one Tradition (I) differs from the other (II) by the de­nial of the existence of so-called inner behavior. The adherence to a subjective ontology is by no means tied to a subjective epistemology which the methodology of Max Weber so clearly shows - in the most excellent interpretation of H. H. Bruun, Science, Values and Politics in Max Weber's Methodology (1972). In the orthodox Marxist Tradition the subjective dimension of rea­lity is regarded as an epiphenomenon and all knowledge including science is socially determi­ned. The distinction between the Explanatory Tra­dition (II) that transcends behaviorism and the Understanding Tradition concerns, of course, the cognitive status of a faculty called Empathy. True, there is a variety of Marxist positions along the two dimensions in the 2 x 2 Table. Moreover, Bang's argument includes a model type distinction; he states: "This intransigant discussion concern­ing the subject matter of the social sciences -which takes place externally as well as internally between the two traditions at the matascientific level - at last winds up with the different products

Table 1. Meta-scientific traditions

•—.^Ontology E p i s t e r n o ^ ^ - ^ ^ logy Objective Subjective

Objective The Explanatory

the Explanatory Tradition I

Tradition II the Understanding

Tradition

Subjective the Marxist Tradition

the Critical Tradition

202 Jan-Erik Lane

at the scientific level, comprising any general un­derstanding which can respond to challenge and questions. Depending on which world-view func­tion they perform and which meta-subject they explore, these products appear in three different explanatory models. Those derived from theory construction/behavior appear as causalistic mo­dels, those derived from theory construction/so­cial facts appear as quasi-teleological models, and finally those derived from organization of enligh­tenment/social definitions appear as teleological models." (Bang 1981:4).

I will argue that Bang's way of making the mo­del distinction between causal models and teleo­logical ones is inadequate and that an understand­ing of the scope of a systems model for the in­terpretation of political data depends upon how that model distinction is conceived. I believe that the Bang argument can be developed still further once it has been clarified how the systems model relates to the distinction between causal and te­leological models - a task to which I now turn, a task that requires an examination of a few basic problems in systems analysis. Bang makes a num­ber of assertions about these fundamental prob­lems in systems analysis; his manner of treating them asks for an elaborate comment. Bang states: "The incorporation of the teleological model with­in the materialistic framework of systems analysis springs in this way from a methodological holism which, according to Easton, stands between pure individualistic reductionism and philosophical ho­lism - i e a holism which asserts both the need to take account of the special properties found when individuals interact as an organized group and the need to reduce behavior to observable individual interactions." (Bang 1981:11-12). The clarity in the debate may be enhanced if these topics lumped together by Bang are treated on their own terms: systems analysis, holism and the nature of a teleological model. These notes contain an outline for such a disquisition.

§1 . For the purpose of stating what is characteristic of a systems model it is expedient to make a dis­tinction between specific systems theories like the Eastonian one and the General Systems Theory. Specific systems theories deal with limited pheno­mena like the political system, the social system or the personality as as system, whereas the Ge­neral Systems Theory - GST - contains a theory or an outline of a theory for all systems. Since

GST is incoplete in several ways and its cognitive status is unclear it should be analyzed separately, because criticism against GST is not necessarily relevant to specific systems theories.

§2. There is a set of works or concepts that usually occur in specific systems theories. The vocabulary of systems theories has a certain degree of co­herence and there is a set of minimum words common to the theories, models and approaches labeled "systems analysis".

§ 3 .

The word "system" is sometimes given a narrow definition, sometimes a wide definition. The Hall-Fagen standard definition of "system" is vague but not ambiguous: DF1. A system is a set of objects (elements) with

relations between the objects and between their properties (Hall & Fagen 1956:18).

From the Hall-Fagen definition follows that it is difficult to point out something in the social world that is not a system. Most social phenomena may be regarded as composed of parts, and some re­lation between these, e g time and space relations, can always be identified. The wide range of the word "system" according to DF1 is due to the fact that no restrictions have been placed on the three key words in the definition. A narrow de­finition of "system" introduces qualifications on object, relation or property.

§4. Some theorists have stated criteria which would sort out trivial from non-trivial systems, thus in­troducing narrow definitions of "system". Ex­amples of such criteria are: (a) constancy in time: only such relations between the parts as persist over time would be accepted as constitutive of a system (Parsons 1965:36). (b) causal relations: only such relations as can be called "causal" can be accepted (Parsons 1966).

Two objections may be made against such cri­teria. It is difficult to create unambiguous criteria of the kind suggested: how long should a relation persist in order to bind together a non-trivial sys­tem?'The use of the criterion causal relation would result in an exclusion of social systems usually considered non-trivial, e g systems where only re­lations of space are valid between the parts. As Easton empasized in Framework for Political Ana-

The Systems model . . . 203

lysis (1965) the classification of systems into in­teresting and uninteresting ones depends on the problem and cannot be made a priori; if the pro­blem is to study the social system that is composed of the interaction of two or several actors, then certain relations are obviously interesting like po­wer, prestige, approval, etc whereas a time pro­perty may be uninteresting. The definition of "system" is necessarily vague in order not to ex­clude a priori applications that could prove inte­resting.

§5. Certain kinds of systems and certain properties of these systems are of focal interest within social science systems theories; other kinds of systems and properties are not considered relevant. Sys­tems theory deals with empirical systems which are described by means of theoretical systems. Sometimes the latter are also called "models". An example'of an empirical system is the political system and examples of theoretical systems are Easton's and Almond's theories of the political system.

§6. A distinction may be introduced between abstract systems and concrete system. This distinction di­vides the set of empirical systems into two sub­sets. An example of a concrete system is a person whereas the aspects of that person that enter into an interaction system constituting a part of a social system belong to an abstract system (Levy 1952: 35-36). Whereas a concrete system exists apart from other such systems, abstract systems are units that come about as a result of various degrees of abstraction from the concrete; such entities are not capable of a separate existence apart from these concrete phenomena. Systems theories deal with abstract systems, called "analytical sys­tems". A social system is an abstract aspect of concrete systems, viz persons, and a political sys­tem is an aspect of social systems.

§7. Static systems are often distinguished from dy­namic systems: DF2. A static system is a system whose relations

are only relations of space. DF3. A dynamic system is a system into which

relations of time enter. From these definitions it follows that static sys­

tems are unchanged over time. An example of a dynamic system is the Swedish political system. The systems dealt with in systems theories are usually dynamic systems.

§8. The concept of environment is crucial in the Eas-tonian systems model. DF4. An environment is the set of objects that

has an input-relation or an output-relation or both to a system.

The application of the concept of environment depends on the choice of system. If ej is an en­vironment, it is an environment for a system a,, a 2 or a 3. A set of objects can constitute a system as well as an environment but not so that ej = aj. What belongs to a system and what belongs to its environment is settled by the problem. In some cases it may be instrumental to divide the set of objects into two classes: a system and an environment. If one wants to describe the struc­ture of a political system at a certain time, it may not be necessary to introduce the concept of en­vironment, whereas this may be required in a change analysis, which is of course the case in the Eastonian systems analysis. It is possible to divide the empirical, abstract and dynamic sys­tems into closed and open systems. DF5. A closed system is a system that has no

environment.

§9. A social system can be part of another more com­prehensive social system just as a (para) political system can be part of a more comprehensive po­litical system. On the basis of the distinction be­tween system-subsystem a distinction between two different levels of analysis can be introduced: macrosopic-microscopic. Macro analysis focusses on properties of wholeness, whereas micro ana­lysis aims at an analysis of the properties of sub­systems. The property of wholeness describes the extent of dependence between the component parts of a system: a system is characterized by wholeness to the extent to which a change in any part brings about a change in every other part (Hall & Fagen 1956: 21). Some systems are highly sensitive to change in their constituent parts like Third World political systems; polyarchies on the contrary may withstand major changes in some

2

204 Jan-Erik Lane

parts without suffering alterations in other parts. A modern society can be classified as more de­pendent in the system sense than an agrarian so­ciety.

§10. Progressive segregation occurs in a system to the extent that the degree of differentiation is in­creasing (Hall & Fagen 1956:22). Progressive se­gregation in a system takes place when the system passes through a transition from a living state to a dead one. Take e g the division of a political system: the different roles that were earlier con­nected in a hierarchical network are separated from each other to be integrated into two new hierarchical networks (Germany in 1945). Another example of this property is growth. "Growth" re­fers to the state of differentiation of a system into subsystems which are more or less functionally specific. A favorite social science example is the division of labor. When classifying societies some measure of segregation like the degree of diffe­rentiation of roles often forms a basis of division for distinguishing between primitive and modern societies.

§11. Progressive systematization takes place in a sys­tem in so far as there is an increase in the amount of wholeness (Hall & Fagen 1956:22). Progressive systematization occurs when systems are integra­ted into new systems and when new relations be­tween the different parts of a system are created. A typical example is the integration in a society of different collectivities. A political system may undergo progressive segregation at the same time as it passes through a progressive systematization. The transition from a primitive polity to a modern one means that out of a number of more or less autonomous collectivities with a low degreee of division of labor a differentiation of roles takes place. Every collectivity is not responsible for all the functions, but a division is brought about. With regard to the performance of functions pro­gressive segregation occurres. At the same time a progressive systematization takes place as the functionally specific subsystems become depen­dent on each other with respect to the conditions for the carrying out of functions. Progressive se­gregation and systematization may constitute ac­tion parameters in the political system, which is of course the case in the Eastonian systems model.

§12. Centralization is a systems property. A system is centralized in proportion to one part being a dominant part, i e a small change in that part brings about a change in all the other parts, and this part is independent of changes in the other parts (Hall & Fagen 1956:22). The property of cen­tralization may occur in connection with the pro­gressive segregation and the progressive sys­tematization of a political system; the property may function as a causal factor between progres­sive segregation and progressive systematization, eg national roles in a society.

§13.

The systems concept of compatibility or harmony stands for a relation between a system and its environment or between a system and another system or between subsystems within a system. It is a property for varying degrees of coexistence and well integrated behavior, summarizing causal relations between different structures and between different functions. The Parsonian concepts of universal imperatives and structural imperatives may be regarded as an application of this concept (Parsons 1966:26-36, 177-180). A fundamental idea of Parsons is that a society is a more or less centralized system that can persist over time only if it is compatible with the environment and if the structure of the dominant subsystem is com­patible with the structure of the other subsystems. With the help of the concept of compatibility con­cepts like conflict and reciprocity can be intro­duced into a systems analysis.

§14. A subset of systems has the capacity of responding to stimuli from the environment with stimuli di­rectly to the environment. Systems can be clas­sified according to the degree of responsiveness. DF6. A feedback system is a system with the

property of introducing outcomes of one or several output-states as input-states.

Whereas responsiveness only implies the capacity for direct external response, feedback means the capacity for indirect external response, i e the capa­city for external response to stimuli from the en­vironment through a process in which outcomes of an external response are mirrored back to the system. Two different kinds of feedback responses can be distinguished.

The systems model . . . 205

DF7. A negative feedback system is a feedback system in which the effects of external sti­muli are counteracted by system responses.

DF8. A positive feedback system is a feedback system in which the effects of external sti­muli are strengthened by system responses.

§15.

In the Eastonian analysis political systems are re­garded as included in the set of stable systems. DF9. A stable system is a system in which a state

is maintained over time. The state that the variables of a stable system maintain or tend to maintain is called "equili­brium" or "homeostas". Two kinds of equilibria are possible, static equilibrium and dynamic equi­librium. The word "homeostas" is sometimes re­served for dynamic equilibria. Adaptive stable sys­tems are dynamic equilibrium systems; they are described .by means of negative feedback; closed stable systems are characterized by stationary equilibrium. Open non-adaptive systems and po­sitive feedback systems cannot be stable systems. It is important to distinguish between the state of the system that is maintained over time and the values of the variables that maintain the state. The variables of a system with dynamic equili­brium can vary in value compared with each other and yet the values of the variables produce a con­dition, which brings about the equilibrium state of the system. Much of the difficulties with the Eastonian systems model stems from the fact that it is not clear how political systems are to be in­terpreted as stable system, what persistence is and what factors effect system persistence.

§16. Political systems seem to belong to a special subset of systems, so-called ergodic systems. DF10. An ergodic system is a system whose state

of equilibrium is not dependent on the ini­tial state of the system; a change in the initial state does not necessarily bring about a change in the equilibrium.

Ergodic systems are the adaptive stable systems which have the property of equifmality. Systems with equifmality can attain dynamic equilibria from a number of different initial conditions, and systems can be classified according to the degree of causal dependence on the initial state.

§17. The scientific value of the taxonomy of the sys­tems concepts and systems types introduced can be appraised only instrumentally depending on if its employment results in theoretically interes­ting propositions and fruitful classification of em­pirical systems. This question should be sharply separated from the evaluation of General Systems Theory. The foremost advocate of a general sy-tems theory is Bertalanffy in his General Systems Theory (1968). In Bertalanffy's articles there are many different ideas; a criticism leveled against some of these is not necessarily relevant with re­spect to other ideas. Some can be adopted without contradiction; others must be rejected.

§18. It is necessary to sort out the various components entering into what Bertalanffy calls the "General Systems Theory":

(a) Specific systems theories of certain properties of the organism (Bertalanffy: 155-185, 120-138).

From a methodological point of view these theories are non-controversial: specific systems theories should be judged according to prevalent criteria: on the one hand the fact that they are empirically scientific theories, on the other the evidence elicited through tests.

(b) A specific systems theory about the diffe­rence between two different kinds of systems, clo­sed and open systems (pp. 139-155).

The following applies to closed systems: the first law of thermodynamics, according to which the total amount of energy in a system S is con­stant or dU=dQ-dW (U=energy, Q=heat and W=work); the second law of termodynamics, ac­cording to which the amount of free energy in S can only decrease or d s>0 for every change in the state of S (s=entropy). The state of equi­librium in a closed system is a static equilibrium, and has properties like: does not perform work, probable, chaotic state; one - one determinism according to classical mechanics. The following applies to open systems: the first but not the se­cond law of termodynamics: of open systems it is true that the amount of free energy can increase or ds=des+dis (es=external entropy and is=in-ternal entropy). As the. externally introduced en­tropy may be negative, the total amount of en­tropy may diminish. This applies to dynamic sys­tems of equilibrium, which are characterized by: improbable, organized and capable of performing work; many - one determinism: equifmality,

206 Jan-Erik Lane

overshoot and false start. What Bertalanffy calls "closed" and "open" systems have been defined above as a closed, stable system and an open, stable and ergodic system respectively. According to K Boulding all kinds of social systems belong to the set of open systems (Boulding 1956).

(c) General Systems Theory (GST).

§19. Bertalanffy points out that GST is not a complete theory (Bertalanffy:55). The object of the theory is as follows: "Its subject matter is formulation of principles that are valid for 'systems' in general, whatever the nature of their component elements and the relations or 'forces' between them." (p. 37). There would be some significant differences between specific systems theories and a complete general systems theory. The basic problem in GST is how to go about finding the underlying struc­ture of all systems? The fundamental idea in Ber­talanffy is that the invariances that characterize systems are to be found in isomorphisms between different laws; such laws are called "homologies": "The isomorphism found in different realms is based on the existence of general system prin­ciples, of a more or less well-developed 'general system theory'." (p. 84). The first task for the de­velopment of a general systems theory is the col­lection and classification of isomorphisms be­tween laws. Two laws are isomorphic if notwith­standing the fact that what they state something about are different objects and different relations, an arrangement of the objects of the two systems can be made which preserves these relations (Kap­lan 1964:185). Laws may have the same structure, e g an exponential function. On the basis of the existence of isomorphisms between laws Bertal­anffy puts forward two theses which can be re­garded as fundamental to GST.

(a) The ontological thesis reads: "(The homo­logy of system characteristics)... is a formal cor­respondence founded in reality in as much as it can be considered as constituted of 'systems' of whatever kind." (Bertalanffy:85). An isomorph­ism confirms the ontological thesis only if the fact that the same mathematical relation applies in both laws also means that there is a common pattern underlying the laws and that this pattern is characteristic of nature. That need not be the case. The ontological thesis is erroneous. It is not possible to deduce empirical properties from me­rely mathematical properties.

(b) The episternological thesis states that the

existence of isomorphisms shows " . . . the fact that certain laws of nature can be arrived at not only on the basis of experience, but also in a purely formal way." (Bertalanffy:62-ó3). This is incor­rect. It is possible to state that two laws are is­omorphic. That is an ex post statement. But ex ante this is impossible. How is it possible to know between what laws there is an isomorphism, i e what ontological structure similar formulas have in common?

§20. The task for GST is to elaborate a system of isomorphisms. The theory would in effect consist of propositions about mathematical structures and GST would become a purely mathematical discip­line investigating all possible kinds of functional relations. It would then not be an empirical theory, since as a mathematical theory it would not fulfil the condition of falsifiability (Popper). Nor does it possess any explanatory or predictive power -it cannot be tested empirically. C. Hempel states: "Thus, e g, in the study of structural similarities between the laws governing different classes of phenomena, system theory could not predict 'new' isomorphisms: it would have to wait for empirical research to establish general laws for a new field, and it could then examine those laws for structural similarities with previously accepted ones." (Hempel 1951:315). A general empirical systems theory, i e a theory of all kinds of systems, constructed on the basis of the two theses spe­cified, the ontological and the episternological one, is a logical impossibility.

§21. The ideal for GST is an axiomatic theory, for­malized in mathematical terminology. Bertalanffy gives an account of an attempt at a theory with such a structure. A number of the systems con­cepts introduced above in a non-formal style are described mathematically with the aid of diffe­rential equations (Bertalanffy.60-80). However, a mathematical description of empirical concepts does not result in an empirical theory of systems, as axioms or propositions are lacking. The trans­lation into mathematical language may increase the exactness of the concepts, but it does not result in knowledge about reality.

§22. Bang fails to recognize the importance of disting-

The Systems mode) . . . 207

uishing clearly between specific systems theories and a general systems theory, GST. Whereas the former are empirically scientific theories and can be studied from several points of view like cog­nitive structure and confirmatory evidence, GST is not an empirical theory. GST can be a tool when constructing theories; it has the same status as paradigms and taxonomies, namely that of an ap­proach. I now turn to two salient problems in sys­tems analysis, reductionism versus holism and the nature of teleological explanation - two subjects which Bang treats in an unsatisfactory manner.

§23.

Systems theories and GST in particular developed from philosophical notions like vitalism, the te­leological idea of the universe, and ideas of emer­gence. All of them are opposed to empiricism, atomism and the mechanistic idea of the universe. It is often maintained as Bang does that the latter ideas are characterized by reductionism and the former by holism. It is possible to define and ana­lyse the two doctrines starting from the following proposition:

(HS) "The Whole is more than the Sum of its Parts"

Holism maintains that the proposition (HS) is true, whereas reductionism rejects (HS). However, the contrast between holism and reductionism is not necessarily contradictory. It may be contrary, i e a third possibility exists. Such a possibility has been put forward by E. Nagel in his analysis of the holistic thesis and the reductionistic thesis (Nagel 1966, Nagel 1969).

§24. (HS) is either trivially true or trivially false de­pending on how the key words are defined. The following example can be used: Let K' = the or­dered set ( 1 , 3 , 5 , . . . , 2,4,6 . . ) , Kj = the ordered set ( 1 , 3, 5 . . ) , K 2 = the ordered set (2, 4, 6...), K = the set of natural numbers and K is not an ordered set, K, = the set of all odd natural numbers and the set is not ordered, K 2 = the set of all even natural numbers and the set is not ordered. (HS) can be reformulated into:

(PI) The set K' is a whole that is the sum of the parts K, and K 2;

(P2) The set K' is a whole is the sum of the parts K.J and K2.

The first proposition is logically false whereas the second one is logically true. Provided that there is a whole consisting of ordered parts, i e

parts and relations, and "sum" means addition of the parts regardless of the relations, then (HS) is trivially true. But if "sum" means addition of ordered parts, (HS) is trivially false. A way of sa­ving the holistic thesis is to maintain that a part A of a whole is something else than A, when it is not part of a whole. The whole determines the properties of A. However, such a proposition is self-contradictory, as A is first identified with A and then A is said to be distinct from A (Nagel 1966).

§25. Whether a whole like a political system can be reduced to the sum of its parts or not is an em­pirical question and cannot be decided a priori. Only in relation to a theory of the whole concerned can the problem of reduction be decided and this is an empirical question. And it is an empirical question whether there is a theory that fulfils the conditions of reduction or not: (a) The words for the whole and the words for the ordered parts must be extentionally equivalent; (b) The laws for the whole must be deduced from the laws for the ordered parts (Nagel 1969). The controversy can thus be solved by making a distinction be­tween a whole that is the sum of its parts and a whole that is not the sum of its parts. The dis­tinction is relative to a theory and thus connected with the development of knowledge.

§26. The word "function" occurs in systems theories. It is often used in different senses in one and the same theory and it often has different mean­ings in various theories (Nagel 1961: 522-526). Causal explanations are characterized by the fact that their propositions often contain words like "cause", "effect", "necessary condition" and "sufficient condition". In functional explanations propositions are used in which words like "in or­der to" and "purpose" occur. An example of a functional proposition is a political science pro­position like: Political parties have interest aggre­gation as a function (Almond & Powell 1966).

§27. Firstly, it is important to distinguish between two kinds of functional explanations on the basis of two different kinds of objects for such explana­tions and two different terminologies that dis­tinguish between these two kinds of explanations.

208 Jan-Erik Lane

On the one hand, the word is used about func­tional explanations in the means-end-terminology of an action performed by a human actor. On the other hand, the word is used of biological systems, personality systems, social systems and cultural systems. The object of such an explanation is some kind of system, and the key terminology is not the means-end-terminology, but structural-functional terminology. The term "functional ex­planation" is often used as a synonym for "te-leological explanation". "Teleological model" may refer to an explanation of an action performed by an actor in terms of means and end. Or "te­leological model" may refer to an explanation of a system in terms of structure and function. Bang does not observe that the distinction between the two types of teleological models is of vital im­portance, as the explanation of an actor's means-end-oriented action is different from the struc­tural-functional explanation of a system.

§28.

Secondly, it is sometimes maintained, as Bang does, that functional explanations constitute a special kind of explanation. A functional expla­nation is said to imply an explanation of systems in terms of inverse causality: a unit in a system is explained by reference to its effects. Suppose that y is an effect of x; the question "why does x occur" is answered in a functional explanation with: "x occurs because y". Functinal explanation has been accused of implying vitalism or inde-terminism. It is believed that a functional expla­nation of a social system implies that the system is conceived to be analogous with the conscious behavior of a human actor and to the system are ascribed properties like purpose, consciousness and action. These correlates of functional expla­nation are not necessarily part of that type of ex­planation. Vitalism, indeterminism or inverse causality can be distinguished from functional ex­planations without the specific properties of the latter disappearing. The distinction between func­tional and causal explanations involves two se­parate problems: (a) The semantic problem: a dis­tinction between causal models and functional models; (b) The ontological problem: a distinction between functional systems and non-functional systems.

§29. The semantic problem concerns the relation be­tween causal and teleological models. There are

two possibilities: (a) Propositions in causal models do not state the same as propositions in functional models. They are true in relation to different con­ditions and are neither equivalent nor intertrans­latable; (b) Propositions in causal models and pro­positions in functional models can be equivalent and consequently intertranslatable.

Strong arguments have been put forward in fa­vor of thesis (b) (Wimsatt 1972). The difference between a causal model and a functional model need not be a difference between an approach that operates with concepts for causal relations and one that operates with other concepts. The dis­tinction could be of a terminological nature.

§30.

A functional model is often said, as Bang does, to be adequate for certain kinds of systems, so-called "functional systems", whereas causal mo­dels are considered to be applicable to non-func­tional systems. What is meant by a "functinal system" as opposed to a "non-functional" sys­tem?

§31. Newton's analysis of the solar system can be taken as an example of a non-functional system. Let S be the solar system and Xj, X 2 , . . . X n the properties of S that are relevant in the analysis, the state variables, and SD be as set of values of the different state variables at a time t, a so-called state description. Newton's theory then shows that if the value of each state variable at the time t 0 is independent of the value of the other state variables at t 0 , SDt[ can be determined with the aid of a function from SDt 0.

§32. Let S' be a functional system, e g an organism, G the state in S' and A, B, C and D the state variables of the system, where A & B & C are endogenous and D is exogenous, namely the en­vironment. Let K A , K B , and K D be a set of possible values for the state variables A, B etc, and let SD be a combination of values from all these sets (A x, B C z , D w). Let K s be the set of all possible SD s. Some of these SD's are G-states, i e such states in (A x , B , Cv D w ) as cause G. Let Ky be the set of such SD's and Ky consist of four subsets K A Y , K B Y , K^y and K D Y - The greater Ky is, the higher is the degree of plasticity or stability. Such a system, is deterministic if the

The Systems model . . . 209

values of the state variables at t 0 , i e S D 0 , are independent of each other, and if the values of the state variables at t j , i e S D , , are a function of SDQ. And such a system is a functional system if and only if (i) there are relations between the state variables in S D 0 and S D , , i e a series of func­tions f,, f2, . . . fn, of such a nature that a change in the value of one variable at t 0 is counteracted by a change in the value of another variable at t j , so that (ii) there is a series of functions g,, g 2 , . . . g n , of such a nature that S D 0 and S D , cause G. A functional system, i ea system in which the state variables and the state G fulfil the conditions (iH'i) above, is a self-regulating system.

§33. It is an empirical question if and to what extent a system is a self-regulating system. Whether a political system is a self-regulating system or not depends on the variations in the values of the state variables, i e K A Y , K B Y , K ^ y and K D Y and whether the relations OH") above are fulfilled. Empirically it may be that political systems are more or less self-regulating, or that some political systems have this property but not all political systems. A theory is about a functional system if and only if in these theories are specified:

(a) S. (b) G in S, i e the state in S the necessary and/or sufficient conditions of which are sought, (c) The state variables in S.

(d) The values of the various state variables: empirically possible values, the G-values. (e) The set of possible SD's. ( 0 The set of SD that brings about G, KQ. (g) The set of functions f,, f2, . . . fn and g,, g 2 , . . . g n .

In some theories of the social and the political systems these systems are assumed to be func­tional systems. Such an assumption is adequate, if and only if the theory contains a description of entities a-g above. If the theory cannot specify these entities the theory as a theory of a functional system is inadequate; these systems may be func­tional systems, as only knowledge of relevant facts is missing. In other theories of social and political systems these are not assumed to be functional systems, which does not prevent them from being such systems, but the theory does not state this. In such theories certain states in a system are also studied, but it is not assumed that these states are self-regulating states, i e it is not assumed that these states are maintained over time according to conditions (i)—(ii) above (§ 32). The distinctions in § 28 may be described in a 2 x 2 Table.

§34.

In his Sociologists, Economists and Democracy (1970) B. Barry states: "The discussion in this book is deliberately limited in scope so as to permit the examination of some ideas in depth. It is fo­cused on two kinds of theory, which differ in a

Table 2. Functional model and functional system

> v Semantics

Ontology ^ ^ - ^ ^ Causal models Functional models

Non-functional systems

Newton's analysis of the solar system

Levy's analysis of society. Almond's analysis of the political system.

Functional systems

Somerhoff's and Bertalanffy's analyses of the organi sm

Parsons'analysis of the social system. Easton's analysis of the political system

2 1 0 Jan-Erik Lane

number of ways. One type is axiomatic, economic, mechanical, mathematical; the other is discursive, sociological, organismic, literary. They differ in the questions they ask and the answers they give. They rest on different ideas about the most fruitful ways of abstracting from the overwhelming par­ticularities of 'reality', and on different assump­tions about the nature of the scientific enterprise itself." (Barry 1970:3-4). By "the sociological ap­proach" Barry refers to systems theory. There is no a priori reason why systems theory should have a different internal structure from other kinds of theory. In some cases a systems approach has been tied to the kind of properties Barry associates with the sociological approach. However, such an as­sociation is accidental and is due to the way in which some scholars have employed the systems approach. Systems theory may use an axiomatic and mathematical style of presentation, it is not by necessity organismic or discursive, and in ge­neral it does not imply a philosophy of science peculiar to itself. In his article in Statsvetenskaplig Tidskrift H. Bang has shown how the debate con­cerning the major systems theoretical work in po­litical science - the Eastonian systems model -may be organized in terms of three perspectives on the systems model. The Bang analysis is il­luminating until he begins to treat the basic met­hodological problems in systems analysis. I argue that Bang's treatment of these problems requires clarification. Moreover, I argue that the Eastonian systems model is a specific systems theory that makes a few basic assumptions about the nature of the political system, the political system being regarded as a functional system. According to my view Easton has not: (a) stated clearly how the political system is to be modeled as a functional system, according to the criteria in § 33; (b) pre­sented confirmatory evidence för the hypothesis that the political system is a functional system.

Of course, if these two tasks, (a) & (b), could be accomplished it would mean a Kuhnian re­volution in political science. Bang's ideas are in­teresting not only in a Scandinavian context, be­cause it seems as if the use of the systems model in political science is by no means exhausted as long as problems (a) & (b) remain to be solved. Let us look somewhat more closely at the im­plications of the criteria of a functional system set forth in § 33.

§35.

Two different kinds of social science systems

theories will be analyzed below; both are examples of functional analysis, which provides an oppor­tunity to state more exactly what a functional ap­proach implies. One of the theories deals with a non-functional system; the other contains an analysis of a system as a functional system. Methodological differences between the theories will be highlighted and an attempt is made to decide whether the theory which looks upon the social system as a functional system fulfils the conditions of adequacy specified above for such a theory (§ 33).

§36.

The Almond and Powell functional analysis treats the political system as a non-functional system; their analysis is an example of structural-func­tional theory. The structure of the analysis is as follows: (a) Definition of the system S to be stu­died, (b) Definition of the environment of S. (c) Establishment of the functional conditions (what is done) for the persistence of S (the functional requisites), (d) Establishment of the structural conditions (how what must be done is done) for the persistence of S (the structural requisites), (e) Clas­sification of the subsystems which are conditions for the occurrence of the conditions in (c) and (d). The classification will show the similarities and differences between the various subsystems that fulfil one or several functions. The Almond-Powell theory consists of two different compo­nents, a classification of poltical systems and a theory of the transformation or development of political systems.

§37.

The taxonomy into which political systems are classified is deduced by means of the structural functional method, (a) S = the political system, (b) Environment. No exact definition of the values for this parameter is given. The political system is an open adaptive system (i e has input and out­put), but the relation between the system and its environment is not stated explicitly (Almond & Powell: 19-21). (c) Functions (pp. 27-30): (I) The capacity function: (i) regulative function, (ii) ex­tractive function, (iii) distributive function, (iv) responsive function. These functions exhaust what must be done by the system relative to the environment. As the concepts demand and sup­port specify relations between the system and its environment, there are corresponding kinds of

The Systems model . . . 211

demands (i-iv) and corresponding kinds of sup­port (i-iv). Similarly, outputs (actions and deci­sions) can be divided into (i-iv). (II) The conver­sion function: (i) interest articulation, (ii) interest aggregation, (iii) rule formulation, (iv) rule appli­cation, (v) rule adjustment, (vi) communication. (Ill) The adaptation function & system mainte­nance function: (i) socialization function, (ii) re­cruitment function, (d) Every function is fulfilled through one or more structures. The different structures can be classified in terms of analytical properties: the degree of differentiation, the degree of specificity, the value of the dicotomous pro­perties of structures (Parsons' pattern variables) (Parsons 1966, Levy, pp. 222-24G). (e) Enume­ration of the various subsystems that are condi­tions for the categories in (c); derivation of the structure of S by identification of properties of the subsystems. The theory determines the struc­ture of S by means of two properties: the degree of role differentiation and the degreee of auton­omy of subsystems in S. (0 Political culture: every political system has a political culture. This va­riable may be introduced in a structural-functional analysis as an abstract aspect of structures. The theory classifies political cultures by the degree of secularization.

§38.

A taxonomy of political systems is introduced on the basis of three analytical properties: degree of role differentiation, degree of autonomy of subsys­tems and degree of secularization, which variables have three possible values: high - medium - low. Thus there are 3 3 types of systems, i e 27 possible types. The theory states that two properties, de­gree of role differentiation and degree of secu­larization covary, which implies that the taxon­omy contains 3 2 possible types. Political systems are classified into this taxonomy (Almond & Po­well, p. 308).

§39.

It should be emphasized that the classification di­vides political systems on the basis of structural properties. Political systems could also be com­pared functionally, i e with respect to the degree of performance of the three functions (I—I'll). If the theory contained only a structural classifica­tion, the word "theory" would hardly be inade­quate. A structural-functional analysis is suitable not only for classification purposes, but it can also be used for the generation of hypotheses. The

theory contains some fundamental propositions that are to be regarded as theoretical generaliza­tions (pp. 322-325): (LI) The higher the degree of structural differentiation and secularization, the higher the degree of capacity. (L2) The higher the degree of structural differentiation and seculari­zation, the higher the degree of responsive cap­acity if and only if the autonomy is high. (L3) A political system persists, only if there are struc­tures for state building, nation building, partici­pation and distribution. (L4) A high capacity oc­curs, only if there is an interest group system or party system or a functional equivalent. Of these lawlike propositions (LI) - (L4), (LI) is the central one. (L4) can be considered a special case of (LI). (L2) can be looked upon as a qualification of (LI) and can enter into (LI) through a minor emen­dation. (LI) has great empirical content or great deductive power, i e by the addition of other hy­potheses a number of propositions can be dedu­ced. These propositions (LI)-(L4) constitute only a fragment of a theory. The theory contains a variety of generalizations about political systems. However, these are not integrated with the pro­positions (LI) - (L4).

§40.

Parsons' functional analysis treats the social sys­tem as a Functional System; in Parsons' writings there are a variety of functional theories about different aspects of the social reality. Parsons has at least three different theories about the social system: (1) System 37: the means-end-termino­logy is used to analyze the basic units from which a system is constructed (Parsons 1968). (2) System 51: structural-functional terminology fTGTA & TSS). (3) System 56: structural-functional termi­nology with emphasis on function (Parsons 1965, Parsons & Smelser 1956). The main features of the theory from 1951 will be outlined below. Since the system 51 is a functional theory of the system conceived as a functional system, the purpose of the analysis is to try to establish whether the theo­ry fulfills the conditions of adequacy for a theory of a functional system as introduced above.

§41.

The structure of the system 51 is as follows: (1) S = a society. (2) Type of analysis = functional analysis of the society as a functional system. (3) G in S = sufficient complementarity of roles and complexes of role in order that collective and in­dividual goals may be realized. The description

212 Jan-Erik Lane

of G in S is vague. In a society G can occur to a varying extent. The state G is threatened by conflict in the society, but G can occur even if there are conflicts within subsystems in S; G can be a state in transformation. (4) State variables in S = (A) role-content; (B) allocation; (c) inte­gration; (D) environment; (E) mechanisms in S relative to its environment.

§42. Role content (TGTA, pp. 208-221, TSS, pp. 113-150). (a) Possible values of the variable A (KA): The possible values K A are established by an analysis of the analytical properties of two kinds or roles, instrumental and expressive. The possible values K A are the types of roles which are combinations of three dichotomous properties of roles: (i) discipline - affectivity, (ii) specific -diffuse, (iii) universalism - particularism. The number of combinations is 2 3 and the different types of roles are shown in Diagram 3 (TGTA, figure 14).

The different categories (1) - (8) are ideal types. In reality most roles do not contain only three of the six possible values; only marginally is a role merely either universalistic or particularistic. Still less can sets of roles, i e societies, be classified in one category only. Different subsystems in a society can be characterized with the aid of (1) - (8) depending on the dominant type of role. A society can be described in an analogous way depending on the subsystem that dominates. Central here is the concept of the dominant type of role, (b) G-values of the variable A (K A G ) : The G-values of the variable A are certain combina­tions of instrumental and expressive roles: those role combinations in which there occurs discipline on the one hand and affectivity on the other, and where the former dominates: (5) or (6) + (1) or

(2), i e four G-values; (4) + (1) or (2), i e two G-values. Combinations with (7) are marginal.

§43. Allocation (TGTA, pp. 197-202). (a) Possible va­lues of the variable B (Kg): The possible values K B are a combination of the values of three var­iables: (i) what is allocated: role, actor, utility, re­ward; (ii) how is what allocated: ascriptively - ac­hievement oriented; (iii) mechanism of allocation: tradition of competition. In an allocation there occurs at least one of the values of the variable (i), exactly one value of the variable (ii) and at least one of (iii). Particularly important is the al-locaton of certain general utilities, money and po­wer, (b) G-values of the variable B (K^j): Certain values of K B are not G-values. This is true of ascriptive allocation through competition, of ac­hievement oriented allocation through tradition, and generally of every allocation value that is not compatible with other allocation values, eg al­location of money by ascription if it is combined with allocation of power on the basis of achie­vement.

§44. Integration (TGTA, pp. 202-204). (a) Possible va­lues of the variable C (Kç): The possible values Kç are a combination of two variables, collective value orientation and sanction role. The values of the first variable are self-orientàtion versus col­lective orientation. The number of sanction roles, i e roles oriented towards the maintenance of col­lective role expectations, can vary in number and differentiation. The occurrence of collective orien­tation in connection with sanction roles is called institution and institutionalization makes up Kç.. (b) G-values of the variable C ( K ^ ) : It holds good

Diagram 3. Possible types of roles in the system 51 Particularism Universalism

Specific Diffuse Specific Diffuse

Discipline (3) (4) (5) (6)

Affectivity (1) (2) (8) (7)

The Systems model . . . 213

that the extreme values, no institutionalization (total self-orientation) and total collective orien­tation are not G-values.

§45.

Environment (TGTA, pp. 226-227, TPS, pp. 26-36, 326-330, 384-390). (a) Possible values of the variable D (Kp): The possible values of this variable are a combination of the values of two variables: (i) endogeneous: personality systems, cultural systyems, subsystems; (ii) exogenous: physical environment, other social systems, (b) G-values of the variable D (KQQ): Certain values in K D are not G-values: (i) endogenous: instability of personality systems and subsystems, conflict between subsystems and inconsistency of cultural systems; (ii) exogenous: those values of the phy­sical variable that render social coexistence im­possible, conflict between S and other social sys­tems.

§46.

Mechanisms in S in relation to its environment (TGTA, pp. 223-230, TSS, pp. 201-243). (a) Pos­sible values of the variable E (KE): The possible values K E are a combination of the values of two variables, socialization and social control: (i) so­cialization: learning mechanisms, consensus me­chanisms, (ii) social control mechanisms: mani­pulation of rewards and deprivations, isolation, reintegration, (b) The G-values of the variable E (K E G ) : The G-values are specified by a number of propositions about the conditions for deviant behavior. The existence of a lack of socialization or a lack of social control respectively and of low-degree of socialization and social control are not G-values.

§47.

State descriptions: Possible state descriptions (Ks) (TGTA, pp. 221-223, TSS, chapters 3 & 4). The possible number of state descriptions is at the same time the number of possible social struc­tures. These are defined by means of the com­binations of the values of the five pattern-variab­les: (i) discipline - affectivity, (ii) specificity -diffuseness, (iii) universalism - particularism, (iv) ascriptive - achievement orientation, (v) self-orientation - collective orientation. Theoretically, there are 2 5 social structures. If a society does not have one social structure, but a structure for each subsystem, the analysis of the possible struc­

tures of a socity becomes more complex, but not different in principle.

§48.

G state descriptions (KQ) (TGTA, pp. 221-223, TSS, chapter 5). The system 51 contains a number of propositions about the state descriptions (SD) in which G occurs. The basis is a fundamental axiom: every subsystem in a society has a struc­ture and the structures of a society cannot vary at random in relation to each other. As K Q is the union of the subsets K A G , KgQ, K^-Q, KQQ, K E G , the G-values of the state variables specified above must be fulfilled. Every SD, i e SD 0, SD,, . . . SD n , must consist of these G-values and be ordered with regard to the G-values. A number of propositions try to solve the problem of spe­cifying KQ: "One of the formemost of these is the problem of the discovery and explanation of certain empirical clusters among the formally pos­sible clusters." (TGTA, p. 222). The system 51 sta­tes some propositions about clusters fo structures in a society, i e combinations of structures for each subsystem in a society, which are a SD G or part of a SD G (TSS, pp. 151-200): (LI) In every society the following entities exist: family, instrumental roles; expressive roles; the territory, violence and power; integration of values (TSS, pp. 153-167). (L2) In a society the dominant structure is com­patible with the other structures if G occurs (TSS, pp. 167-180). (L3) If in an society a dominant structure is universalistic, it is also achievement-and discipline-oriented. (L4) If in an society a do­minant structure is affective, it is also particu­laristic and ascriptive (L5) If in a society a do­minant structure is universalistic and achieve­ment-oriented, there is no other dominant struc­ture that is particularistic and ascriptive. These lawlike statements (LI) - (L5) do not exhaust the part of the theory relevant for the specification of K Q , but they state certain essential relations in the system 51.

§49.

The set of functions: f,, f2, . . . f2 and g,, g 2 , . . . g 2 . The system 51 does not specify any of these functions, though it speaks of self-regulating functions in connection with certain subsystems (TSS, chapters 7 & 8). The theory assumes that the society is a functional system, but the two sets of functions that must be specified in order that the term "functional system" may be applied are not specified.

214 Jan-Erik Lane

§50. The system 51 does not satisfy the conditions of adequacy for a theory about a functional system that were laid down above (§ 33). The specification of KQ is incomplete, and the sets f,, f2,... f2

and gj , g 2 , . . . g n are missing. The system does not show that a society is a functional system. However, the theory contains several necessary components of such a theory: "Our own analysis is thus very far from a classification of actual struc­tural types of social systems. But it does present, we feel, a systematic approach to the problem, which is capable of further development into the very heart of substantive theory. It delineates all the principal components - the elements of orien­tation and the functional problems which it will be necessary to incorporate into such a classifi­cation - and works out some of their relations to each other." {TGTA, p. 223).

§51. There is a variety of political theory types like systems theory, the sociological approach, rational choice theory, the actor approach, public choice, formal poltical theory, the structural-functionalist approach, collective behavior, game theory and so on. Whereas systems theory used to be fashionable during the sixties rational and collec­tive choice approaches have become en vogue du­ring the seventies. These ways of modelling po­litical data do not offer contradictory alternatives for theoretical interpretation. In fact, they may very well complement each other. It is not a daring assumption to predict that we will see a revival of the systems model in political science during the eighties. Probably that revival of the relevance of the systems model will mean a renaissance of the Eastonian systems analysis: even if Bang is only partly right when reinterpreting Easton's system analysis, then maybe Danish political scientists like G. Sjoblom, L. Lundquist and E. Rasmussen are not that old-fashioned, after all.

Literature Almond, G. A. & Bingham Powell, G., (1966) Com­

parative Potties. Boston & Toronto: Little, Brown & Company.

Bang, H., (1981) "The Conflict between the Traditions' in Statsvetenskaplig Tidskrift 1981:1.

Barry, B., (1970) Sociologists, Economists and Democracy. London: Collier-MacMillan.

Bertalanffy, von., (1968) General Systems Theory. New York: George Braziller.

Boulding, K. (1956) "General Systems Theory: The Ske­leton of Science" in Management Science, 2.

Easton, D. , (1965a) A Framework for Political Analysis. Engelwood Cliffs: Prentice-Hall.

Easton, D. , (1965b) A Systems Analysis of Political Life. New York: John Wiley & Sons.

Emery, F. E. (ed) (1969) Systems Thinking. Harmond-sworth. Penguin.

Hall, A. D. & Fagen, R. E., (1956) "Definition of Sys­tem" in General Systems Yearbook I.

Hempel, C. G., (1951) "General Systems Theory and the Unity of Schience" in Human Bilogy 23.

Levy, M , (1952) The Structure of Society. Princeton: Princeton University Press.

Lundquist, L., (1972) Means and Goals of Political De­centralization. Lund: Studentlitteratur.

Nagel, E., (1961) The Structure of Science. London: Ro-utledge & Kegan Paul.

Nagel, E., (1966) "On the Statement 'The Whole is More Than the Sum of Its Parts' " in Lazarsfeld, P. & Ro­senberg, M. (eds) The Language of Social Research. New York: Free Press.

Nagel, E., (1969) "The Meaning of Reduction in the Natural Sciences" in Danto, A. C. & Morgenbesser, A. (eds) Philosophy of Science. New York: Meridian Books.

Nagel, E., (1969) "A Formalization of Functionalism" in Emery (ed).

Parsons, T. (1968) The Structure of Social Action. New York: Free Press. First published in 1937.

Parsons, T. & Shils, E. A. , (1965) "Values, Motives and Systems of Action" in Parsons & Shils (eds), Towards a General Theory of Action. New York: Harper & Row. First published in 1951 (TGTA).

Parsons, T., (1964) The Social System. New York: Free Press. First published in 1951 (TSS).

Parsons, T., (1965) "An Outline of the Social System" in Parsons, T. & Shils, E., Naegele, K. D: & Pitts, J. R. (eds), Theories of Society. New York: Free Press.

Parsons, T. & Smelser, N. }.,(\956)Economy and Society. London: Routledge & Kegan Paul.

Popper, K. R., (1965) Conjectures and Refutations. New York: Harper Torchbooks.

Rasmussen, E.,(i91l-12)Komparativ Politik, Second ed. 1-2. Copenhagen: Gyldendal.

Sjöblom, G. (1968) Party Strategies in a Multiparty Sys­tem. Lund: Studentlitteratur.

Somerhoff, G., (1969) "The Abstract Characteristics of Living Systems" in Emery (ed).

Wimsatt, W. C , (1972) "Teleology and the Logical Structure of Function Statements" in Studies in His­tory and Philosophy of Science 3.